Function normalize 

pub fn normalize(input: &[u8]) -> String

Normalize note body bytes before hashing.

Rules (§5.3):

1. newlines normalized to \n
2. trailing whitespace stripped from each line
3. text Unicode-NFC normalized
4. exactly one trailing newline

Input is treated as UTF-8; replacement chars are used for invalid bytes so hashing never fails on a partially-written file.